JPH0594001A - Manufacture of submicron device - Google Patents

Abstract

PURPOSE: To make it possible to easily obtain a phase mask which imparts the values of plural phase delays to pattern plotting radiations by forming openings within a mask layer and heating a material enclosing these openings. CONSTITUTION: A resist layer is formed over the entire surface of a substrate 10 and is patterned and plotted by electron beam writing using the writing of, for example, an electron beam exposure system, by which the resist layer 11 is selectively removed to afford consumed shape holes 12, columns 13 and rectangular line projections 14 and to expose the substrate surface 15. Next, the functional masking layer 16 of averaged five thicknesses is constituted by heating up the resist material 11 constituting the projections 13 and 14 by enclosing the openings 12 or allowing the material to flow by another means. Namely, the resist layer consisting of the five thicknesses composed of the initially deposited regions 17; the regions 18 obtd. from the flow for filling the openings 12; the regions 19 obtainable from the openings enclosing the projections 14; the regions 20 and 21 where the flow is limited in design; and the non-packed regions 22 is formed.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION This invention relates to the manufacture of small size devices, such as integrated circuits using submicron design rules, and the equipment / tools used in this manufacture. The broad but major impetus relates to lithographic writing, ie the use of phase masks to improve image quality. Whether based on currently used writing energy, eg, energy in the near ultraviolet spectrum, or shorter wavelengths, eg, wavelengths in the deep ultraviolet or x-ray spectra. The improvement of lithographic drawing expands these ranges and enables further miniaturization. The manufacture of large scale integrated circuits, ie electronic as well as optical and hybrid integrated circuits, manufactured according to submicron designs is of interest here.

[0002]

2. Description of the Related Art The story of large-scale accumulation from the beginning to the present is well known. The development of current 1-2 megabit chips, which are manufactured according to design rules of 1 μm or a little lower, does not mean an ultimate product. Lithographic drawing is expected to play a major role in the future as in the past. Modern device manufacturing relies on the use of near-ultraviolet radiation (eg wavelength, λ = 3650Å, mercury I-line radiation). The energetic efforts towards the next generation of devices require much shorter wavelength radiation (“deep U”).
V "spectrum, eg wavelength, λ = 2480Å, radiation within the krypton fluoride excimer laser line). Future work towards smaller design rules is in the x-ray spectrum. We are considering the use of electromagnetic energy, or similarly short wavelength accelerated electron radiation.

Competing efforts have been expended in expanding the capabilities of the UV writing radiation currently in use. M. D. Levenson et al., IE
EE transaction. Electron device (IEEE T
rans. Electron Devices), Vol. ED-29 (1
2), page 1828 (1982), and as reviewed by the New York Times paper of December 12, 1990, the loss due to edge resolution design rule constraints is "phase mask ( phase mask), ie, the use of a mask designed to provide a relative phase shift of the radiation when transmitted through selected mask areas. This shock is double. That is, (1) when applied to normal device manufacturing involving the use of opaque shaped masks (eg, chrome on glass) and (2) when applied to manufacturing involving the use of transparent masks that eliminate the need for opaque mask shapes. However, in the latter case, an image of dark lines is used due to the interference between transparent mask areas of different phase delays. In either case, the use of a phase mask allows the design rules to be extended beyond what would normally be beyond the capabilities of the particular wavelength used, but this extension is at the shape edge. Resolution-constrained radiation (diffraction-scatte)
red, resolution-limiting radiation) phase cancellation. In both cases, adjacent to the edge drawing mask area or as an integral part of this, 18
A 0 ° phase shift region is provided.

Phase masking appears to be promised in the light of traditional business considerations. This use promises the use of next generation devices using current equipment and processing methods. Certainly, efforts will continue in this direction, as the cost of replacement equipment (although not yet on the market today) and retraining of employees can be avoided.

[0005] According to a widespread view, UV-based processes use phase masking to provide effective UV writing capabilities up to design rules of 0.3-0.25 μm and below, and in some cases generally. It is expected to extend to the range of 0.2 μm or below, which is considered to exceed. If this proves correct,
Device fabrication by use of x-rays (independent of proximity or projection) and accelerated electron radiation (independent of beam writing or masking) will be delayed until the end of the century.

The limit of lithographic resolution varies according to the classical equation (1):

[0007]

[Equation 1]

Where λ = appropriate unit, eg μ
is the wavelength of the writing radiation in m and NA is the aperture of the optical system. The resolution is based on the required shape-space contrast. K ₁ is a constant that depends on the details of the imaging system and the characteristics of the drawing process, eg, the developing process. The value of 0.7-0.8 is (0.8-1.0μ
This is a characteristic value in the latest manufacturing (of LSI having a design rule of m).

A 180 ° phase mask treatment for a given wavelength / etch contrast reduces K ₁ to ≈0.5 levels (allowing device design according to 0.4 μm design rules), or in some cases. Can be explained in the form of a reduction of K _{1 to} a level of ≈0.3 which allows a quarter micron shape.

[0010]

SUMMARY OF THE INVENTION The present invention relates to device manufacturing with at least one level of patterning that relies on image transfer from the mask to the device during manufacturing.
Intrinsic propulsion was designed to provide multiple values of the phase delay and amplitude of the writing radiation for the purpose of improving image quality, for example, improving the resolution of the edges of the imaged shape. With the use of masks. The edge resolution is increased independent of the feature size, but this enhancement is particularly effective for small features, that is, small features with small spacing. Therefore, the main object of the present invention is to obtain a shape that follows the submicron design rule, that is,
It generally entails the manufacture of features contained within integrated circuits. A related approach is directed to the use of smaller device design rules than would normally be achievable with a particular writing radiation. A design approach within the scope of the present invention is the co-filed US patent application SN / 07/67
3614, which is incorporated herein by reference. This co-pending patent application discloses a circuit design approach that has significant implications when implemented in accordance with the present invention. Without being limited to this, an example of an important effect relates to the image shape produced by interference. This example includes line termination and line branching, but to this end the advantage of using different values of the phase delay is exploited. In addition to this, the spacing of geometries often compromised due to the phase mask approach is greatly enhanced by creating closely spaced geometries at different phase delay values of the pair of writing radiations.

The provision of a 180 ° phase mask as described in the literature (see above) from a co-pending patent application is
Although the resolution is greatly improved, it can be seen that there is room for further improvement. The approach described in this document is based on phase cancellation based on the assumption that the energies to be canceled are of uniform phase. This assumption does not take into account the phase delay introduced during imaging, for example through the proximity effect of closely spaced features. Such a variation from the nominally constant phase becomes more significant as the circuit design rule becomes smaller. The actual pattern wavefronts of the submicron design rule devices considered are characterized by varying, usually continuum, phase variations. Undesired radiation emanating from the mask, eg diffractive edge scattered radiation, has such a phase variation, and measures in the sense of phase cancellation should take this into account. Total cancellation continues to require 180 ° out-of-phase energy, provided that it is (uniformly with respect to the nominal phase of the virtual uniform pattern wavefront ignoring proximity and other perturbation effects). Fixed 180
This means 180 ° out of phase with the locally scattered energy to be reduced (rather than a phase delay of °).

From a mask fabrication perspective, the above objective is to provide at least two levels of phase variation, ideally having a series of values (eg, 5 ° or 10 ° apart). ) It is required to have a means for providing a phase continuum or near continuum. Variations are conceivable by changing the thickness and / or refractive index of the mask layer material through which the patterning energy is transmitted, or alternatively by changing them for a mask operating in the surface reflection mode. The phase retardation of the present invention may be provided in the reflective mask in various ways, either separately or in combination. Local variations in the required length of the phase path are in the form of variations in the removal of free reflective surfaces; variations in the penetration of incident radiation in the transparent layer to the reflective surface; variations in the refractive index within these layers. Will be realized. These masks, whether in transmissive mode or reflective mode, are designed to provide three or more levels of phase delay for unwanted radiation cancellation / reduction, where Phase delay mask (Multi Phas
e delay masks (MPM).

The manufacture of MPMs by extension of the procedure used for the selective removal / retention of unchanged thickness (full width) mask layer regions carried out in 180 ° phase masking as described in the literature It is avoided in the invention. Providing a region of additional value for the phase delay by this means is costly because it requires an additional layer of Cusk as well as an additional lithographic step associated with it, which statistically reduces device yield.

The present invention provides a device manufacturing method by the use of an easily prepared mask that provides a plurality of phase values, and even this continuum, without the need for an additional lithographic writing step. These masks are considered a two-step process regardless of whether they operate in transmissive mode or reflective mode:
1. 1. quantitative or "digital" correction of defined local mask areas, and Physical averaging of these modifications,
Manufactured by a process consisting of. In an embodiment where the phase delay only results in a thin layer,
Step 1 consists of the local removal of the layer material to obtain the opening material, and step 2 is the opening by the induced flow of the peripheral layer material (here "expendable aperturing").
es) ”) backfilling. This results in a layer thickness variation that gives the required phase variation (in addition to the pattern information needed at that level). A layer is obtained, the openings having various shapes and sizes, and in one embodiment in the form of circular holes of uniform thickness with a distribution / density that provides the required material removal. The shape of the aperture may include slits, columns or areas that define the ground, etc., as suggested by the possibly device geometry.Variations give the addition of material rather than removal of material, and always This is followed by the "averaging" step.
In this averaging step, the height of the patterned additional material is flowed in this example.

One particularly important aspect of the present invention is that the wear shape can be backfilled to the extent required (while avoiding retention of meaningful device functionality in the final pattern), while on the other hand, the device pattern drawing shape. And the ability to hold non-device shapes included on the mask as compensation information. A particularly important example of the latter is a grating to reduce unwanted brightness variations in the final pattern. The convenience of approximately four power dependence of the flow velocity of the backfill material on the distance traveled facilitates and preserves the separation of these two (ie, the shape to be filled and the shape to be retained). A distinction is made between "large" size features to be filled and "small" size features to be filled. However, it is considered that the same degree of filling occurs with respect to the shape to be retained, and as a result, the scale of surface removal is reduced. The teachings of the present invention, in some cases, minimize this effect by overdesigning the shape to be retained to the extent required to achieve the required removal dimensions after flow. Termination of flow steps is accomplished by simple timing or, if important, real-time monitoring. As an example of the latter, a grating designed to have excess peak-to-vessel height is monitored to identify the endpoint of the process in relation to the required diffraction scale. Other variations are described in the detailed description section.

Once the rheological understanding is obtained, the mask design rules can be naturally understood. In summary, backfilling is
Intellectually, this too is considered to consist of two steps. The initial flow responds to the high free surface energy of the substrate exposed during pattern writing. For a properly selected material that provides the requisite small wetting angle, this requirement is met by a thin layer of flowing material (about 100 or less layer thickness). Mutual flows in response to surface tension considerations are strongly dependent on the dimensions of the void. That is, it has sufficient reliance to be able to backfill a consumable opening without significant optical impact on the void (non-consumable opening) to be retained. These two considerations are the basis for both material selection and design geometry.

A first aspect of the invention is the projection-reduction i using UV writing radiation.
Expected in relation to maging). However, in principle,
It is applicable to contact (proximity) printing as well as 1: 1 projection printing, and also to the use of shorter wavelength electromagnetic radiation (eg radiation in the x-ray spectrum).

[0018]

[Basic Considerations] Fortunately, previous studies have considered many of the parameters needed to implement the process of the present invention. For example, L. E. Stillwagon and R.W. G. The Journal of Applied Physics (J.
Appl.Phys.), Vol. 63 (11), (1988 June 6)
(March 1 issue), published in page 5251, "Topographical substrate leveling fundamentals (Fundamentals o
f Topographic Substrate Leveling) ”. The addition of other work to the information contained in this article constitutes the basis for various parameters that may be useful in accordance with the teachings of the present invention.

The experimentally verified observations presented in this reference are noteworthy and, to some extent, are the basis for a detailed discussion of all aspects of the invention. It should be noted that the fourth power depends on the distance traversed by the flowing material. Translated for the purposes of the present invention, this dependency serves as a base for holding larger sized features. That is,
For example, the amount of material that is brought to the center of the 10 μm smallest dimension feature to be retained is 10 ⁻⁴ of the 1 μm expendable feature amount. This resulting variability can be compensated for even if it is sufficient for process-differentiation, or with additional processing, for example, uniform exposure processing to cause selective etch removal. There is also.

Material Selection Manufacturing in accordance with the present invention provides for backfilling of consumable openings.
) Accompanied. The composition of the materials involved, for example the composition of the material of the layer in which the openings are to be made, as well as the composition of the "substrate" material which is exposed in making the openings, must first meet the usual requirements based on patterning. All these properties are well known and well documented for most materials. The basic text is well known to those of skill in the art, but the Stillwagon (Stillwagon, cited above)
The direction suggested by the article by Agon et al. reflects one satisfactory direction for resist selection.

An important material property is the backf
low) The simplest approach is explained in connection with FIG. 1, but here the resist layer itself also plays the role of this function. Inevitably the process simplification required in terms of yield is a factor of sufficient commercial choice. Resist materials are well known and many have promising flow properties. Functional separation, which results in relaxed requirements, requires the use of a second layer to transfer the pattern, this flow, and also the final phase-dependent layer (thickness and / or surface). Various materials with poor photochemical properties (relative insensitivity to the writing radiation), but with favorable properties in relation to their subsequent stability in the presence of radioactivity as well as their mechanical properties, for example. Exists. One group of these materials is "spin-on" glass, a commercially available low melting glass that can be used to form layers in conventional spinning processes.

In general, "thickness should be averaged (thic
The desirable properties in the context of "kness-averaged)" materials are obvious. These properties include, for example, chemical properties that give the required durability / reactivity both during manufacture and in subsequent use; , Physical properties that ensure the required dimensional stability depending on temperature / environment; optical properties that meet the usual mask requirements, etc. Additional requirements are imposed on manufacturing in accordance with the teachings of the present invention. The requirements for
(A) Both the critical category of submicron LSIs, eg, the minimum requirements to achieve the parameters encountered in the fabrication of LSIs with design rules of 0.5 μm or less, and (b) the smaller and larger dimensions. And the following processing and operational considerations:
That is, the interface between the substrate and the backfill material that guarantees a contact wetting angle that is low enough to withstand or require an acceptable temperature rise; both essential requirements and economic considerations. It is convenient to discuss in the context of ranges that take into account energy values etc.

Note that the discussion here is primarily about masks. The cost of manufacturing the mask is of course important, but it is not so important in terms of the manufacturing cost including the final product. That is, the increase in mask cost can be easily recovered by a small standardized product cost improvement.

In many embodiments, where the patterned material is flowed, eg, by promoting flowability, and possibly by chemical phase change, a major requirement is flow characteristics. Perhaps the viscosity, which decreases with increasing temperature, must be low enough to meet the process time goal. The surface tension of the backfill material in the liquid state must also be sufficient to meet the time requirement in relation to the degree of surface flatness achieved. Both of these properties are temperature dependent.

The initial wetting of bare surfaces by lithographic processing (eg directly by exposure as by evaporation or indirectly by subsequent development) is aided by a small Ett angle. In common situations this is not a problem. Typical organic resist materials can be heated without adversely affecting the temperature levels, resulting in near zero angles. Generally, angles of 10 ° or more may be used if other circumstances require. It may be necessary to include additional layers selected primarily for the purpose of providing a low wet angle, but these layers contact this underneath the resist or other layers to be backflowed. And placed.

As shown, energy requirements are usually met by backflow and wet layers of less than 100Å are obtained. The requirement of 100Å or less is a consideration for less important optical effects and can be easily compensated for in the mask design. Alternatively,
Such a small thickness of layered material can also be easily removed, for example by exposing the whole to a suitable etchant without significantly affecting the rest of the mask.

In the context of initial wetting followed by backfilling, it is fortunate that the dependence of the flow rate on the opening size serves as an effective basis for a controlled process. The above-cited paper "Basics of Topographic Substrate Leveling" gives an experimental verification of the fourth power dependence on distance obtained as a conclusion regarding flowing materials during backfilling. This particular condition is that both have substrates lying under the same free energy, and both are surrounded by a source material (layer material) of the same thickness, with the same configuration of surface boundaries for the openings. We need a comparison between two apertures that we define. In a precise mathematical solution it is necessary to consider various factors including, for example, the variation of the driving force causing the flow, ie the layer thickness depending on the surface-to-surface slope and the difference in the dimensions of the openings. is there. Nevertheless, the order of magnitude of flow-transport in the majority of backfilling operations is governed by this quadratic dependence.

From the results of the above considerations, it is necessary to define the ratio of the consumable opening size to the non-consumable opening size (eg, the shape opening to be retained as opposed to the consumable opening to be backfilled). . The material parameter that governs backflow is the ratio of surface tension to viscosity. The material to be backflowed is, in the traditional sense, a glassy material. From this point of view, whether glassy or not, a suitable material is a continuous series of subsequent properties in relation to fluctuating temperatures within the subsequent temperature range (within the temperature range in which the flow is carried out). Required to show dependency. Such materials, whether inorganic or organic, have a suitable range for the purpose:
It shows a viscosity that decreases sharply with temperature within a range of maximum temperature corresponding to the retention of viscosity in the range slightly above the softening temperature and required for control.

Appropriate backflow conditions, that is, mainly temperature and time determinations, are made practically. According to this approach, the mask is now heated and held at a temperature sufficient to cause the required backflow, after which the mask is cooled to terminate the backflow. Variations are also effective in the context of the diffraction grating.
Monitoring based on the above is performed by detecting the intensity of the diffracted radiation while using the material. It is convenient to use a laser-generated beam for irradiation, and detection is done using a photodiode. It is unlikely that such a result would require adjustment due to the difference in wavelength associated with the intended writing radiation.
This can be easily calculated if needed.

Page 5252 of the Stilwagon et al. (Cited above) document presents the following "dimensionless time" formula useful for determining termination points on a computational basis. .

[0031]

[Equation 2]

Here, T is a unitless constant obtained from FIG. 13 of the same reference, for example. t is the backflow time. Under the assumption of constant temperature, the actual conditions characterizing this process are close enough to the assumption. Backflow times, which are typically in the range of minutes for these processes, together with cooling, which essentially terminates the backflow within seconds, supports this proposition. h ₀ is the physical driving force for the flow at different heights at the interface between the material to be flowed and the bare substrate to be backfilled. w is the minimum size of the substrate area to be backfilled. (If addressed, it is the minimum dimension of an equal line / spaced grid, where w is the width of the line or space in the region of interest.) Γ is the surface tension, and μ is the viscosity.

As an example, these values are h = 0.5 × 1
0 ⁻⁴ cm; w = 0.25 × 10 ⁻⁴ cm; T (if determined from FIG. 13) = 500 × 10 ⁻⁴ ; t = 3 minutes.
These substitutions into the above equation yield the ratio γ / η = 0.087 cm
/ Sec. give. A typical surface tension of 30 dynes / cm (characteristic of a suitable resist material) gives a viscosity of 345 poise.

Mask Design / Manufacturing Considerations The approach according to the present invention has some requirements for masks. Matters decisive for the functional device pattern are ensured by the invention rather than retained. However, in the design of the shape, the part which is not the functional device part must be taken into consideration. Such non-device geometries (“compensation geometries”) include a grid / checkerboard structure (unfilled and image transfer) to meet consumable aperture and grayscale requirements to be filled in the flow. In the final structure, both of which have a size and a configuration such that they do not remain, are included in the final structure.

The mask generation consists of two phases. That is, a phase in which a mask that produces a particular image on the wafer is designed by specifying the phase shift and amplitude of the wavefront of the light coming in through each point on the mask; and specific to each point Of masks with phase shifts and amplitudes of.

A specific amount of phase and amplitude when exiting the mask
The calculation of the image produced by the light with the cloth is simple
It's a problem. This calculation can be done fairly accurately.
Are described in textbooks on optics. This
Regarding this, for example, Reynolds, Debe
Squirrel (DeVelis), Parent (Parrent), Thompson
(Thompson)Fourier Optics Lecture
(Tutorials in FourierOptics), SPIE optical engineering
Zineering Press Publishing, 1989, page 27
"Physical Optic Notebook (Physical Optic
s Notebook) ”.

However, the inverse problem, one wants a particular pattern on the wafer, and knowing what phase delays and amplitudes introduced by the mask produce that pattern is a more complex problem. There are several different answers that produce almost identical results. Also, it is not possible to get the ideal pattern, one has to be satisfied with the best compromise, and the definition of what is a good compromise is usually a matter of value judgment.

An important category of masks relies on transparent layers consisting of regions of different thickness to produce a particular phase shift of the transmitted or reflected light at individual points. This mask is usually created by the use of an electron beam to expose the electron resist so that the resist is completely removed at one point and remains at another point.
The pattern in the electronic resist is then transferred to another material of suitable physical properties (eg, suitable softening temperature and viscosity) to meet the flow requirements.

It is necessary to discuss related mask dimensions. The discussion is done in the context of future device generations manufactured according to the design rule of ≈0.35 μm (assuming a continuous 5x reduction projector, this results in a minimum of 1.75 μm on the mask. The dimensions are given: 1. The smallest consumable opening on the mask, 0.25 μm, is several times smaller than the 1.75 μm feature and is believed to have a filling rate of 2,401 times the device feature. 2. The gray scale is created by diffracting the light exiting the system The grating and checkerboard pattern that creates the gray scale is (λ = 0.365 μm
, With a diffraction-determined dimension of the order of 0.75 to 1 micron, and a back fill ratio for 0.25 μm consumable features is 81 to 256 times.

The smallest feature to be printed on a wafer, ie the smallest feature according to the allowed design rules, has dimensions of the following order:

[0041]

[Equation 3]

The corresponding size of this feature on the mask now follows the linear reduction factor of the mask-to-wafer transfer system. A typical reduction factor for the projection system currently most commonly used is 5. λ is the wavelength of the writing radiation, NA is the aperture of the optical system, and K ₁
Is a constant.

Therefore, for a reduction factor of 5, the minimum feature size on the mask is a dimension of 2 μm, which corresponds to a feature size of 0.4 μm on the wafer. Mask shape,
Consumable openings and grid lines and pattern shapes are usually
As a result, when viewed on the mask, it will be larger than calculated by the above equation. This is consistent with the teaching of the other application SN / 0776314, mentioned above. Increasing the size of the mask features increases the sharpness of the edges of the features, for example varying phase and phase shifts used to reduce fringing of interference and other incidental artifacts. This is the result of the grayscale enclosing region.

The initial thickness of the layers involved gives the greatest variety, ie, layer thinning.
It is usually one wavelength or an integer thereof so that it is possible to achieve an arbitrary value in the entire range of the phase delay value corresponding to. This initial value is represented by the following equation (4).

[0045]

[Equation 4]

Here, λ is the wavelength of the drawing radiation, and n is the refractive index.

For example, for the mercury I line (λ = 365 μm for a glassy layer with n = 1.6), the initial thickness is about 61 μm. Material removal or flow according to the present invention allows for any required phase delay.

Drawings The drawings shown are possible commercial versions of the invention by way of example. All have a common purpose:
In particular, precisely control the image front as received on the article to be manufactured, with emphasis on the phase angle of the writing radiation within the image front to improve resolution, as described in detail above. They share the purpose of convenient manufacture of the masks used for. The manufacture of the entire mask or the relevant parts according to the invention always consists of two basic steps: The first procedure is a "digital" removal of material in regions that extend throughout the "resist" layer, the removal of this region having a cross-sectional dimension and distribution such that the purpose of step 2 is achieved. Is done as follows. The second procedure is the "averaging" of step variations caused directly or indirectly by digital removal, which is accomplished by a processing step involving or consisting of material flow. . In the simplest embodiment, the flow of material surrounding the removed area itself achieves its overall purpose. In another embodiment, the flow of material around the openings or in the protrusions as a result of resist pattern transfer; or into the lower material; or into the upper subsequently deposited material (by lift-off). Flow; or patterned, optionally providing a flow of material in the form of an eluting material that remains after removal of the flow-filled solvent material.

1 to 3 are examples of the simplest form of the manufacturing process of the present invention. That is, here, digital patterning as well as backflow filling occur within a single initially provided "resist" layer. For these purposes, the term "resist" has its most general meaning, that is, having the selective "resist action" of subsequent processing, eg, transferring any tone of the resist pattern onto it. Patternability, such as opening patterning by removal or retention (possibly supplemented by addition of material) to allow
ty) is defined. The term "resist" also includes materials that are patterned in response to the writing energy itself, for example by evaporation, as well as in response to subsequent or simultaneous development.

FIG. 1 shows a starting body consisting of a substrate 10 and a supported resist layer 11. While not intended to be limiting, this is described in the context of state-of-the-art manufacturing, where the substrate 10 has lateral dimensions of 6 "to 8" and a fractional thickness of inches. All of these are easy to process,
It is required by robustness and other mechanical properties. The materials for both substrate 10 and resist 11 depend on a number of well-known considerations. For example, if the mask is expected to operate in transmissive mode, the substrate must be sufficiently transparent to be masked.
If the mask is operated in reflective mode, for example, x
− When mask drawing the energy of the line spectrum,
The substrate material is chosen only in relation to its mechanical properties.

The starter body of FIG. 1 is then used, for example, in the Electron BeamExposure System.
Writing is used to pattern by electron beam writing. For a general description of this commonly used patterning approach, see L. et al. F. Thompson, C.I. G. Willson
) And M.M. J. AC by Bowden et al.
S Symposium Series
(1983), page 73, "Introduction to Microlithography"
) ”. Alternatively, a "master" mask may be used, which itself is manufactured by the use of electronic beater writing. The image transfer from the master is a patternless radiation, eg U
It involves raster scanning of the V or x-ray spectrum or mask irradiation with flooded electromagnetic radiation.

One important feature of the present invention is that a single drawing step does not require, or usually avoids, duplication within the device functional pattern as well as the final device, or wear or other shape or other compensation. Shaped patterning is achieved. In other words,
Shapes that are not related to device functionality, but are needed for device manufacturing convenience, such as consumable apertures, diffraction gratings to create gray scale, are associated with the same device function pattern in a single step. Is drawn. This feature of the invention, which allows a single mask layer to be used to perform both operations in a single step, avoids the additional yield loss associated with the additional masking step.

The material from which the resist layer 11 is constructed again depends on the intended mask mode (reflection or transmission), and similar relevant properties of the device writing radiation to be masked. From the perspective of mask fabrication according to the present invention, the resist material, whether it is patterned by the writing energy itself or by attendant development, must have additional properties not normally required for the resist. In the example shown, the resist is
Apart from how it is guided once it is aperture patterned, it must have flow characteristics that give it a timely flow to achieve the fill requirements according to the invention.

FIG. 2 shows the structure of FIG. 1 after pattern drawing. Selective removal of material of layer 11 is consumable shape hole 1
2, pillars 13 and rectangular line protrusions 14 are provided to expose the substrate surface 15.

In FIG. 3, the protrusion 12 is surrounded by the opening 12.
An averaged layer thickness is obtained as a result of the flow of resist material comprising 3 and 14 being induced by e.g. elevated temperature or other means. The associated flow produces a resist layer that comprises the following five thicknesses of functional masking layer 16. That is, the first deposited region 17; the region 18 resulting from the flow to substantially fill the openings 12 of equal size / spacing; the region 19 resulting from the flow filling of the rectangular openings surrounding the protrusions 14;
Resultantly designed to cause local averaging everywhere (including thinner regions produced from less layer material held in the form of wider spaced columns that provide flow material) Areas 20 and 2 where flow is restricted to
1; and finally, a resist layer consisting of five thicknesses consisting of regions 22 that are essentially unfilled, that is to say essentially exposed during writing for other patterning. (As described elsewhere, region 22
In fact, a very thin layer of resist material (not shown), for example imposed by the choice of substrate 10 to provide a bare surface of sufficiently high energy to enable the low angle wetting required by the present invention. Supports 100Å thick material needed to meet the energy requirements required. )

For simplicity, FIGS. 1-3 show some of the masks that will be manufactured, but these parts are primarily, but not exclusively, associated with the compensatory patterning of the present invention. As shown, the present invention has the advantage that device pattern writing is obtained by the same mask at the same time as the pattern variations of the present invention (eg, aimed at improving the resolution of the device pattern). Therefore, the broken section shown may only show the portion of the mask under fabrication that contains such additional device pattern information.

As shown, FIGS. 1 to 3 similarly show a transmission mode mask being manufactured (with the exception of resist layer 1).
It is also possible to use it in the reflection mode by suitable selection of the material of 1). Although not shown, the variations imposed by the flow on the resist layer thickness and the corresponding surface measurements, even for transmissive resists, reflect mode by addition of metal or other reflective coating (not shown) such as by evaporation. Also used for. Another difference is that in this case the lower reflection with variable path length and associated phase delay associated with one round trip path through the treated coating (comprising regions 16-20). Depends on layer (not shown).

4-7, the initial resist patterning is transferred to the lower layer, and the thus patterned lower layer is then flowed in the same manner as described in connection with FIG. One variation is shown. My goal is,
It is to separate the required actinic radiation characteristics of the resist from the required mask characteristics of the drawing material. That is, for example, to substitute a relatively stable and durable inorganic glassy material for the final processed resist.

FIG. 4 shows, in sequence, the substrate 30 supporting a layer 31 which finally functions as a mask pattern layer supporting a resist layer 32. In FIG. 5, here again, for example, as a result of drawing and development by electron beam direct writing, the openings 33, the rectangular protrusions 34 and the columns 35.
While the shape is produced in the form of a substrate surface 36, in this case the surface of the layer 31, is exposed.

In FIG. 6, the pattern is transferred to layer 31, shown as layer 31a, resulting in a positive image again composed of holes and protrusions 37, 38 and 39, while surface 40, In the case, the surface of the substrate 30 is exposed. Shape 33, 34, 3
The positive replicas of 5 and 37, 38, and 39 result from the use of positive tone resists that make up layer 32.

In FIG. 7, for example, the flow induced by an increase in temperature is the result in this case, again in the case of the material of the layer 31 of, for example, low-melting glass, the pattern indicated here as 31b. Six distinct regions within the layer are provided. In the order of thickness, these regions are shown as 41, 42, 43, 44, 45 and 46, the latter of which is on the Angstrom scale (eg, as described above) to meet the energy requirements due to exposure of high energy surfaces. The support material 39 is not provided except for the thickness (about 100Å).

FIGS. 8-10 show yet another variation, in which after the digital removal of the resist material, the consumable openings are finally (to result in averaging the layer thickness). Flow filled.

In this example, FIG. 8 shows the mask in production at the level in that of FIG. 2 or 6. As shown, the substrate 50 includes a material 51 (eg, the resist material of layer 11 of FIG. 1 or layer 31 of FIG. 6).
a glassy material of a). The patterned layer 51 includes holes 52, rectangular protrusions 53.
And columns 54. At this stage of manufacturing, additional layer regions 55 were required for uniform deposition conditions across the potential device portion shown. The choice of metal or other reflective material that functions in layer 55 meets the requirements for one type of reflective mask according to the present invention.

In FIG. 9, the lift-off of the material 51 (li
As a result of the ftoff), only the region 55 in contact with the substrate 50 is held. The duplication due to lift-off, as shown, is the shadow tone of significance, the holes 52 are here represented by posts 56, the protrusions 53 are represented by elongated tubs, and the posts 54 are represented by openings 58. .

In FIG. 10, the flow of material under the properties and conditions described in connection with the previous figures results in local thickness averaging, which results in regions of progressively smaller thickness. 59, 60, 61, 62, and finally a region 63 that is substantially naked (eg, a region of thickness in the Angstrom range such that an energetically defined initial flow always occurs).

FIGS. 11 and 12, in common with the other figures shown, illustrate mask fabrication in which the flow provides multiple levels, as well as continuously varying phase delay values, in response to digital changes. FIG. 11 shows a stage in which the topmost layer 65 supported by the substrate 66 is flowed, resulting in a structure such as that shown in the stage of FIG. 3, for example. The flow here is 67, 68, 6
It provides five discrete regions of decreasing thickness, designated as 9, 70 and 71. At this stage, the backflow-filled structure is exposed to vapor 72, eg, metal vapor, under conditions and for a time sufficient to saturate resist layer 65 throughout its entire thickness.

In FIG. 12, the resist host (or matrix) is removed by dissolution or evaporation, for example, under conditions that leave the metal layer 73. This metal layer 73 is characterized by regions 74-78, which have the same relative thickness to the extent that a uniform thickness is achieved with that of regions 67-71, respectively. Under the processing conditions for the device shown, the vapor 72 does not deposit on the free surface 71 and the layer 65
Retained only by the solution in the host material. Under these conditions, region 78 is uncoated.

In the case where the vapor 72 is a metal that is reflective to the device manufacturing writing energy, the product, ie the mask, is reflective and the stepwise steps required for such energy. With different controlled path lengths giving different phase delay values. In many cases, this approach allows control of very thin layered materials, a degree that is not readily achievable by physical flow.

Device Drawing It is useful to consider the operating characteristics of a mask constructed in accordance with the present invention. A major consideration relates to the wavelength of radiation that should be used in device manufacturing. This section describes this aspect and its relevance to the present invention.

The purpose of the first generation phase mask is to obtain higher resolution and energy quality with the writing energy currently used. This includes energies in the visible spectrum, eg 4360Å mercury G lines; energies in the deep UV spectrum, eg 248.
0 Å krypton fluoride excimer laser line or 2540 Å line of mercury; and ultimate shorter wavelength,
For example, the argon fluoride excimer laser 1930Å line is included.

The present invention is not wavelength limited. It can also be used with writing radiation in the x-ray spectrum, as is used, for example, in projection or proximity printing. US Patent Application SN 595,341, filed concurrently on October 10, 1990, incorporated herein by reference, discloses, for example, longer wavelength portions of the x-ray spectrum, eg,
Disclosed and claimed is an effective projection system operating in reduced mode for wavelengths in the range of 40Å to 200Å. It can also be used in x-ray proximity printing, thereby allowing use throughout even shorter portions of this spectrum, including, for example, the 9Å to 18Å range. In new technology, the same principle is
It also applies to masks for x-ray projection lithography in the 0 to 200Å range. Mask fabrication techniques are also envisioned using x-ray projection printing in the 9 to 18 range.

A significant degree of coherence of the writing radiation is required in the projection system to improve the sharpness of the shape edges in phase-delaying cancellation of Essi scattered radiation as well as other attendant resolution-degrading effects. Irradiation coherence refers to both temporal (longitudinal) coherence and lateral or spatial coherence. Temporal coherence is related to the bandwidth of radiation and can be expressed as, for example, the coherence length represented by the wavenumber. In order for meaningful interference to occur, the phase difference between the waves that combine at a particular point must be less than or equal to the coherence length of the radiation. In interest in phase masks,
The phase difference rarely exceeds several wavelengths. In such a situation, the coherence length obtained by using the exemplary light source shown is sufficiently long to cause no problems.

The other surface is space or lateral coherence. It is characterized by a numerical aperture NA and a filling factor σ. For a laser or incoherent source placed at a distant point (far enough to act as a point source), σ = 0 and there is considerable lateral coherence. For larger sources (eg diffuse point sources) the lateral coherence is small,
Gives a larger σ. In normal lithography a low σ gives sharper edges with smaller shapes,
However, it has ringing and interference effects,
Therefore, a somewhat larger value of σ is preferred. Usually, a value of σ = about 0.5 is suitable. In the case of phase masks, ringing is a serious concern, but again a finite value of σ is appropriate, which represents a compromise between ringing and edge definition.

Currently, phase mask systems are designed to work with existing cameras, for example with σ = approximately 0.5. Masks made in accordance with the present invention use such σ values. Depending on the conditions encountered, smaller compromise values may be used. The use of partially coherent light (a finite value of σ) averages out many effects of interference. Full control of both phase and amplitude eliminates the need for such averaging and allows optimization of edge sharpness achievable by using σ = 0.

[Brief description of drawings]

The drawings attached to the present application show, as an example satisfying the requirements of the present invention, sequential steps of manufacturing a mask shape.

1-3: A two-layer body region composed of a substrate and resist is directly shaped by removal (making an opening), which produces a wear image, which later produces the final mask. One structural approach that is flowed to

FIGS. 4-7 are similar to FIGS. 1-3 but show an approach in which material is flowed after the developed depleted image is transferred to the material layer under the resist.

8-10: The ultimate structure is composed of additional material deposited on the already patterned layer, after which
We present an approach in which the structure is lifted off to leave a mask pattern that depends predominantly or completely on the deposited material.

11-12 illustrate, for example, the fabrication of reflective mode masks that rely on metal vapor. Here, after deposition on already defined / backflowed layers, the metal vapor is diffusion saturated in these layer materials, after which the host material is removed so that only the metal remains.

[Explanation of symbols]

 10 Substrate 11 Resist Layer 12 Openings 13 and 14 Protrusion 16 Masking Layer

Claims (23)

[Claims] 1. A masking layer having a masking pattern defined by drawing areas of different characteristics to enable transfer of the transfer pattern to a transfer surface to ultimately obtain the final pattern by means of transfer radiation. In a method of manufacturing a mask involving use, the final pattern has a minimum dimension of less than 1 μm, and the resolution of the transfer pattern is improved by providing a phase shift characteristic to reduce the effect of lowering the resolution. A shift feature is designed to provide a relative phase shift for the portion of the transfer radiation between different regions of the transfer surface, the masking pattern being of two types of information: (a)
Final mask pattern information to be included in the final mask pattern and (b) auxiliary transfer information not necessarily included in the final mask pattern, wherein the masking layer is manufactured from a precursor masking layer, and the precursor masking layer is underside. An opening that creates a digital pattern defined by a precursor opening area extending through the entire thickness of the precursor masking to expose the material of
The pioneer opening area is 30 of the minimum size of the final pattern information.
Manufacturing method having a minimum dimension not greater than%, the digital pattern being processed by a method comprising at least partially filling an opening with a flow of physical material. 2. The manufacturing method according to claim 1, wherein the decrease in resolution is generated at least in part by diffracted scattered radiation, and the phase shift characteristic is a relative phase shift of 180 ° with respect to a part of the transferred radiation. Is designed to give 15% of the minimum dimension of the final pattern information.
A method of manufacture comprising a consumable region of minimal size no greater than, wherein the flow of the material results in a significant filling of the consumable region. 3. The manufacturing method according to claim 2, wherein the included precursor region includes a non-printing region which is larger than the consumable region but small enough to be excluded from the final pattern. 4. The method of claim 3, wherein the non-printing areas are sized and spaced such that the non-printing areas diffract the final pattern transfer radiation to the extent necessary to avoid hitting it on the transfer surface. A manufacturing method comprising: 5. The manufacturing method according to claim 3, wherein the non-printing region is a diffraction grating. 6. The manufacturing method according to claim 1, wherein the final pattern constitutes a functional portion of a discrete mask. 7. The method of claim 1, wherein patterning is initiated by electron beam writing into the resist layer to selectively remove the resist material and create an aperture image. And manufacturing method. 8. The manufacturing method according to claim 7, wherein the opening image includes a consumable region, the resist material near the consumable region is flowed, and the consumable region is filled. 9. The manufacturing method according to claim 7, wherein the opening image is transferred to a lower layer, and the lower layer is processed, resulting in the masking layer. Production method. 10. The manufacturing method according to claim 9,
A method of manufacturing wherein the processing of the lower layer comprises removing material to create an aperture image that includes a consumable region that is filled by flow. 11. The manufacturing method according to claim 10, wherein the lower layer causes a flow due to an increase in temperature and is essentially composed of a glass layer filling the consumable region. . 12. The manufacturing method according to claim 7,
A method of manufacturing characterized in that an upper layer is produced by depositing a material on the opening pattern resist layer. 13. The manufacturing method according to claim 12, wherein the subsequent processing comprises lift-off of the deposited material by removing the resist so that an image defined by the deposited material selectively retained in the opening is produced. A manufacturing method characterized by the above. 14. The manufacturing method according to claim 13, wherein the selectively retained material is flowed so as to fill a region caused by lift-off. 15. The manufacturing method according to claim 8,
A method of manufacture characterized in that the resulting reflow patterned resist layer is exposed to a liquid that acts as a source of solute for the resist layer under conditions that substantially saturate the entire layer. 16. The manufacturing method according to claim 15, wherein the resist is substantially removed so that the patterned region of the solute remains. 17. The manufacturing method according to claim 16, wherein a distributed Bragg reflector is constructed on the patterned layer of the solute. 18. The method of claim 1, wherein the backflows have different thicknesses and associated different surface measurement numbers sufficient to result in a phase difference of at least 10 ° with respect to the transfer radiation. Producing a masking layer consisting of at least three distinct regions of. 19. The method of claim 1, wherein backflow results in a maximum variation of phase difference of at least 10 ° with respect to the transfer radiation within the continuum and varying thickness and associated different surfaces. A method of manufacturing, characterized in that a masking layer consisting of a continuum of a measured number of layer materials is produced. 20. A manufacturing method according to claim 18 or 19, characterized in that it comprises a duplicating step which, when duplicated, produces a separate mask of negative tones. 21. The manufacturing method according to claim 20, wherein the duplication is achieved by hot pressing. 22. The manufacturing method according to claim 20, wherein the duplication is achieved by casting. 23. The method according to claim 18 or 19, characterized in that the mask to be reproduced is modified by selective addition or reduction of material to enhance the surface to be reproduced.